Human in Loop — CXSphere AI Ops
Human in Loop · AI + Human Judgment

Automate with confidence.
Humans decide when it matters.

CXSphere's risk-based Human in Loop system ensures AI handles routine tasks while humans approve high-risk actions — giving you automation without anxiety.

RISK ASSESSMENT ENGINE EVALUATING
42
Risk Score · Auto-Approve Threshold
Configured Thresholds
Low Risk (0-50) AUTO-EXECUTE
Medium Risk (51-75) HUMAN APPROVAL
High Risk (76-100) BLOCK EXECUTION
Risk-Based Controls

AI scores every action.
You set the thresholds.

Every automated action is scored for risk based on business impact, blast radius, and change history. Configure thresholds to define what AI can do autonomously vs. what requires human approval.

🎯
Multi-Factor Risk Scoring

AI evaluates business impact, user count affected, service criticality, change window, and historical success rate to calculate risk score.

⚙️
Configurable Thresholds

Set org-wide or team-specific thresholds. Define different rules for production vs. staging, critical vs. non-critical services.

📊
Approval Workflows

Route high-risk actions to designated approvers — engineers, managers, or on-call SREs.

🔄
Continuous Learning

AI learns from approval patterns and outcomes, refining risk scores over time to reduce false positives and improve accuracy.

Decision Matrix

Three outcomes.
One decision engine.

Based on risk score and configured thresholds, CXSphere automatically executes, requests approval, or blocks actions.

Auto-Execute

Low-risk actions run automatically with zero human intervention. Full audit trail logged for compliance.

Risk Score: 0-50
  • Password resets
  • Cache clearing
  • Log rotation
  • Service restarts (non-prod)
  • Read-only queries
👤
Human Approval

Medium-risk actions require approval from designated humans before execution. Approvers review context and approve/reject.

Risk Score: 51-75
  • Database failover
  • Production deployments
  • Config changes
  • Access grants (sensitive data)
  • Service scaling
🚫
Block Execution

High-risk actions are blocked from automation entirely. These require manual execution by authorized engineers.

Risk Score: 76-100
  • Production data deletion
  • Firewall rule changes
  • Certificate revocation
  • DNS record deletion
  • Root access grants
Performance Metrics

Automation without
anxiety.

Teams using Human in Loop controls see high automation rates while maintaining safety and compliance standards.

75%
Actions auto-executed
12%
Require approval
3%
Blocked (high-risk)
99.2%
Risk score accuracy

Configure your risk thresholds.

See how Human in Loop controls give you automation with confidence.